minerva-lite-data / preparation /program_catalog_builder.py
miller345's picture
Add seeded environment data, provenance, statistics, and dataset card
f88d59f verified
Raw History Blame Contribute Delete
4.29 kB
import json
import pathlib
from datetime import date
import re
import requests
from bs4 import BeautifulSoup
from common.synthetic_catalog import sanitize_program
ROOT = pathlib.Path(__file__).resolve().parents[1]
SEED_PATH = ROOT / "data" / "seed.json"
OUTPUT_PATH = ROOT / "data" / "program_catalog.json"
def _clean(text):
return " ".join((text or "").replace("\xa0", " ").split())
def _extract_summary(tab):
if not tab:
return ""
paragraphs = []
capture = False
for node in tab.find_all(["p", "h2", "h3"], recursive=False):
text = _clean(node.get_text(" ", strip=True))
if not text:
continue
if node.name in {"h2", "h3"} and "program description" in text.lower():
capture = True
continue
if node.name in {"h2", "h3"} and "degree requirements" in text.lower():
break
if capture and node.name == "p":
paragraphs.append(text)
return " ".join(paragraphs)
def _extract_courses(table):
rows = []
for tr in table.find_all("tr"):
classes = tr.get("class") or []
if "bubbledrawer" in classes:
continue
cells = tr.find_all("td")
if len(cells) != 3:
continue
code = _clean(cells[0].get_text(" ", strip=True))
title = _clean(cells[1].get_text(" ", strip=True)).rstrip(".")
title = re.sub(r"\s+\d+$", "", title).rstrip(".")
credits = _clean(cells[2].get_text(" ", strip=True))
if not code or code.lower() == "course":
continue
try:
credit_value = float(credits)
except ValueError:
credit_value = None
rows.append(
{
"code": code,
"title": title,
"credits": credit_value,
}
)
return rows
def _extract_requirement_sections(tab):
if not tab:
return []
sections = []
current = None
pending_note = None
for child in tab.children:
if not getattr(child, "name", None):
continue
if child.name == "h2":
if current:
sections.append(current)
current = {"heading": _clean(child.get_text(" ", strip=True)), "notes": [], "courses": [], "course_groups": []}
pending_note = None
elif child.name == "p" and current:
text = _clean(child.get_text(" ", strip=True))
if text:
current["notes"].append(text)
pending_note = text
elif child.name == "div" and "courselist-wrapper" in (child.get("class") or []) and current:
table = child.find("table", class_="sc_courselist")
if table:
courses = _extract_courses(table)
current["courses"].extend(courses)
current["course_groups"].append({"note": pending_note or "", "courses": courses})
pending_note = None
if current:
sections.append(current)
return sections
def build_catalog():
seed_data = json.loads(SEED_PATH.read_text())
programs = seed_data.get("programs", [])
session = requests.Session()
built = []
for index, program in enumerate(programs, start=1):
response = session.get(program["official_url"], timeout=30)
response.raise_for_status()
soup = BeautifulSoup(response.text, "html.parser")
tabs = soup.select("div.tab_content")
summary = _extract_summary(tabs[1] if len(tabs) > 1 else None)
sections = _extract_requirement_sections(tabs[2] if len(tabs) > 2 else None)
flattened = []
for section in sections:
flattened.extend(section["courses"])
merged = dict(program)
merged["summary"] = summary or program.get("summary", "")
merged["requirements"] = flattened or program.get("requirements", [])
merged["requirement_sections"] = sections
merged["verified_on"] = date.today().isoformat()
built.append(sanitize_program(merged))
if index % 25 == 0:
print(f"Fetched {index}/{len(programs)} programmes")
OUTPUT_PATH.write_text(json.dumps(built, indent=2))
return built
if __name__ == "__main__":
build_catalog()